[sword-devel] New BibleCS Binaries

[Chris Little]

> 1) U+2014, U+2019, U+201C and U+201D show as boxes (used in 
> lots of modules, eg BWE, HNV, Rotherham, WEB, for long dash, 
> single and double open and close
> quotes)

I think these texts are not really in Latin-1/ISO 8859-1.  They're in
Windows Codepage 1252.  Windows shouldn't have any problem showing
U+2014, etc. but since we assume all non UTF-8/SCSU modules are Latin-1,
we're assigning control codes to these characters.  That's because
cp1252 value 0x97 (equivalent to U+2014) just gets turned into U+0097 (a
control code).

> Additionally, half the letters in the Czech versions, 
> N27U4, PolBTP, etc now appear as boxes.

These have similar issues, but using other Windows codepages or (in the
case of N27U4) special fonts.  The fix in all these cases is to remake
the modules using UTF-8 or change the Latin-1 filters to CP1252 filters.
The latter is simpler and won't adversely affect our real Latin-1
modules.

> 2) 'Options' settings are not remembered

No one has looked into this yet.

> 3) With RtoL texts (seen in both BHS and AraLAB), verse 
> numbers are often at the end (left) of one line, and the 
> verse starting on the next line.

We could newline after every verse like other software does.  I don't
care for this solution because I think it's a cheap hack in lieu of a
real BiDi text rendering solution.  We should re-examine the situation
after we implement HTML rendering and ICU BiDi re-ordering.

> 4) With RtoL texts (seen in AraLAB), chevrons should be 
> associated with the RtoL rendering. It looks strange seeing 
> >> then next verse number, then RtoL text, then << where the 
> >> << are supposed to surround the RtoL text. Are RtoL 
> chevrons available? This is probably only a module problem. 
> Same problem with exclamation marks at ends of verses.

This is partly a bug in Windows' incorrect BiDi implementation and
partly a bug in the Windows' BiDi implementation due to reliance on a
pre-3.1 Unicode Standard.  Chevrons weren't considered mirrored
characters prior to 3.1, but now are.  This should all be fixed in 1.5.4
when we use ICU for re-ordering.  

> 5) Still can't turn off showing of cross references in ISV 
> (and cross references are coded as footnotes in TEV)

Still haven't touched it.  I need to figure out how best to markup
cross-references.  I'm leaning towards using ThML for the ISV now.

> Incidentally - with footnotes turned off, TEV has double 
> spaces where footnotes are hidden. I think this is a library 
> problem more than a module problem.

Nope, that's a module problem. :)  It probably has both leading &
trailing spaces inside outside the footnotes tags (like "blah blah
<RF>something about blah<Rf> blah") but one of them should be inside the
footnotes for correct spacing.

Module bugs don't need to be fixed before the 1.5.3 release, just before
we do a new CD pressing.  The exception is the issues with the ISV's
cross references since they will require a possibly significant amount
of new code to support them.

--Chris